Scan a directory for LLM SDK usage, AI frameworks, exposed API tokens, and hardcoded secrets. Returns all findings grouped by type (token, secret, sdk, framework, endpoint, model) with file locations and severity levels.
AI agents call scan_directory to retrieve information from Ai Scanner without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs reconnaissance by analyzing codebases to detect LLM usage patterns, frameworks, and secrets, but does not alter data or trigger external operations. It is fundamentally a read-only analysis tool. The severity is elevated from low to medium because discovered secrets and endpoints in the scan results could inform subsequent malicious actions if misused by an AI agent, creating informational risk.
From the tool's definition Tool scans a directory and 'returns all findings' without modifying any files or systems. The description explicitly indicates it retrieves information: 'scan', 'returns findings'. No create, modify, delete, or execute operations are mentioned.
Attacks that exploit this kind of access
Scan a directory for LLM SDK usage, AI frameworks, exposed API tokens, and hardcoded secrets. Returns all findings grouped by type (token, secret, sdk, framework, endpoint, model) with file locations and severity levels. It is categorised as a Read tool in the Ai Scanner MCP Server, which means it retrieves data without modifying state.
Register the Ai Scanner MCP server in PolicyLayer and add a rule for scan_directory: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Ai Scanner. Nothing to install.
scan_directory is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the scan_directory rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for scan_directory. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
scan_directory is provided by the Ai Scanner MCP server (sky-in-code/ai-scanner-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →